Academic Work
Publications
Finite Element Modelling of Atomic Force Microscopy Imaging on Deformable Surfaces
RSC Soft Matter | In Press | https://doi.org/10.1039/D4SM01084A
Joshua Giblin-Burnham, Yousef Javanmardi, Emad Moeendarbary and Bart Hoogenboom
Abstract
Atomic force microscopy (AFM) provides a three-dimensional topographic representation of a sample surface, at nanometre resolution. Computational simulations can aid the interpretation of such representations, but have mostly been limited to cases where both the AFM probe and the sample are hard and not compressible. In many applications, however, the sample is soft and therefore deformed due to the force exerted by the AFM tip. Here we use finite element modelling (FEM) to study how the measured AFM topography relates to the surface structures of soft and compressible materials. Consistent with previous analytical studies, the measured elastic modulus in AFM is generally found to deviate from the elastic modulus of the sample material. By the analysis of simple surface geometries, the FEM modelling shows how measured mechanical and topographic features in AFM images depend on a combination of tip-sample geometry and indentation of the tip into the sample. Importantly for the interpretation of AFM data, nanoparticles may appear larger or smaller by a factor of two depending on tip size and indentation force; and a higher spatial resolution in AFM images does not necessarily coincide with a more accurate representation of the sample surface. These observations on simple surface geometries also extend to molecular-resolution AFM, as illustrated by comparing FEM results with experimental data acquired on DNA. Taken together, the FEM results provide a framework that aids the interpretation of surface topography and local mechanics as measured by AFM.
Projects
01
Onsagers Variational Principle As Applied To Multiphysics Modelling Of A Lipid Bilayer (w/ Prof Sonia Antoranz Contera and Prof Antoine Jerusalem)
Recent experiments have underscored the active nature of neuronal membranes, revealing an intricate coupling between their mechanical, electrophysiological, and biochemical properties. To address this complexity, we develop a coupled mechano-electrophysiological and biochemical model for neuronal lipid bilayers. Extending the Onsager variational principle to neuronal membranes, we describe the non-equilibrium dynamics of infinitesimal membrane patches. Through minimising the energetic fluxes and dissipative potentials, we can derive the system's equations of motion. Applying a finite difference simulation we provide a framework for further simulational work and experimental comparison.
02
Dynamic Energy Budget (DEB) simulation of shell morphology and flesh growth (w/ Prof Derek Moulton and Alain Goriely)
The escalating effects of climate change, such as rising sea temperatures and ocean acidification, may impact Queen Conchs. Given the cultural and commercial importance of Conch in the Caribbean, modeling the optimal growth conditions has become of paramount importance to Conch hatcheries. This study presents a dynamic growth model for the Queen Conch that incorporates both somatic and shell morphology to provide comprehensive predictions of shell thickness and overall morphology. Building upon the net production approach, our model integrates a geometric description of the accretive growth process.
03
Watching Wiggling Biomolecules with Atomic Force
Microscopy (w/ Prof Bart Hoogenboom)
Atomic Force Microscopy (AFM) is a versatile three-dimensional topographic technique implementing a mechanical probe to raster-scan and image sample surfaces. However, there are limited computational recreations of AFM imaging, and the area could benefit from greater tools to aid in interpreting surface characteristics. Consequently, this research presents novel computational modelling of AFM imaging using Finite Element Modelling (FEM). These simulations show the viability of the FEM approach in reproducing the AFM dynamics and provide a wealth of extensions to be explored.
04
MAPs Summer Studentship: CONTUR as applied to Gildener-Weinberg Higgs Bosons (w/ Prof Jonathan Butterworth)
A further update to Gildener-Weinberg Higgs Bosons (Summer 2022) was made using the data available in Rivet 3.1.6 and the Contur 2.4 release candidate. My analysis further refined the constraints using new data, as well as some improved treatment of error correlations in CONTUR. The statistical exclusion now shows the limits are much stronger than the initial study, with the whole model being disfavoured at one sigma over the region analysed.
05
CMMP Summer Studentship:
Conquest as applied to Zirconium
Diselenide (ZrSe2) Defects (w/ Prof David Bowler)
The goal of the project is to simulate the structure and STM appearance of various proposed defects with in Zirconium Diselenide (ZrSe2) using density functional theory (DFT) and the CONQUEST DFT code. Existing collaborations with University of Geneva, on a similar material (TiSe2), serves as the starting point for this investigations and alongside knowledge of the crystalline growth conditions of ZrSe2 has give various predictions of the defect structures. As well as investigating the structure of defects in ZrSe2, this project also has enable us to test the effectiveness and accuracy of the CONQUEST basis set for STM simulations.
06
Ant Colony Optimization Algorithms as applied to path minimisation in Graph Theory
The analogous behaviour of ant colonies as cooperative, algorithmic agents and computational heuristic algorithms is adapted as an approach to solve combinatorial optimisation problems. An ant colony optimisation algorithm is implemented for solving path optimisation problems. The pheromone-based communication of biological ants is used as the predominant paradigm and the main characteristics of this model are positive feedback and the use of a constructive greedy heuristic. We apply this methodology to both a simple, two path Minimal Spanning Tree Problem (MST) and the classical Traveling Salesman Problem (TSP).
07
Application of Machine Learning and a Convolutional Neural Network as a Neutrino Event Classifier
We look to replace current neutrino event classification methodologies with Convolutional Neural Networks (CNNs) which have a wide breadth of application in feature learning and categorisation problems and, therefore, playing a pivotal role in image recognition and analysis. We apply these technics to identification of particle interactions in high energy neutrino physics. Using the core concepts of CNNs and deep learning, we built and trained networks using feature extraction and machine learning algorithm, for identifying NOvA-like neutrino interactions based on visual topology.
08
An Experimental Investigation into an Alternative Transparent Conducting Oxide for Application in Industry
Transparent conducting oxides (TCOs) are a unique subset of semiconductors that exhibit both electrical conductivity and optical transmittance. We outline the background of the materials, including their vast industrial and commercial applications; conduct a theoretical analysis of the concepts underpinning the physical properties of TCOs; and, subsequently an experimental study into the structure and band gap of scandium doped zinc oxide (ZnO:Sc) was conducted in UCL’s Chemistry Department. Polycrystalline powders were sintered and studied by X-ray diffraction and UV-vis measurements.
09
Study of an Oscillating Mechanical System Driven into Resonance
Many systems follow simple harmonic motion when driven at their resonant frequency, the oscillations are fortified creating larger amplitudes. This is often dangerous in real world systems, so energy is removed from the system via dampening. In this study resonance was observed in a torsional pendulum, mechanical system. The system was placed under forced oscillation, in order to measure how amplitude and phase of resonance vary with angular frequency, subject to different damping torque from silicone oil with a viscosity of 100 cP, at depths of (5.0±0.5) mm and (20.0±0.5) mm. The experiment resulted in inconclusive results.
Other Code
Alongside my reports within my degree I have developed my computational abilities, learning to code in both Python and Mathematica; I have model various systems from spin interactions in atoms, thermal diffusion, and wave propagation through media that can be viewed on my GitHub page below. Computational modelling and analysis has by far become one of my favourite areas of physics as I find computational work deeply satisfying and fascinating; I’m hopeful to gain further experience in the area and prospectively looking towards computational work when considering a career in research.